Artificial Intelligence-Based Biometric Authentication Systems for Facial Recognition and Identification

Biometric authentication methods are the focus of this study, which also delves into the topic of facial recognition technology. The study aims to address existing issues in face recognition systems, such as accuracy, robustness, and user-friendliness, by recognising the significant impact of artificial intelligence (AI) and deep learning, specifically convolutional neural networks (CNNs). The study used the Olivetti Dataset obtained from Kaggle, which consists of 400 facial photos belonging to 40 people. The research follows a methodical approach that includes data collecting, preparation, and the partitioning of the dataset into training and testing sets. A main goal is to combine hybrid CNN models—LR-XGB-CNN, LR-LGBM-CNN, and LR-CBC-CNN—in order to improve accuracy and resilience. The comparative research demonstrates that the LR-LGBM-CNN model surpasses its predecessors, attaining higher levels of precision, accuracy, recall, and F1-score. Notably, it achieves an amazing accuracy rate of 87%. This study makes a noteworthy addition to the advancement of biometric identification technology by overcoming the limits of existing face recognition systems. It also promotes enhancements in confidentiality and user experience in various situations.